Hi,
I was a believer.
Until I read Eva's posts.
Now I am inclined to not believe until I am capable of testing/measuring the outcome (a bit like Thomas & his finger).
It appears that oscillation can be promoted and HF garbage sent to the output if bypassing is done incorrectly, or not at all if Eva is to be believed.
Go find his/her posts and extensive test results/graphs/diagrams.
Then make up your own mind.
I was a believer.
Until I read Eva's posts.
Now I am inclined to not believe until I am capable of testing/measuring the outcome (a bit like Thomas & his finger).
It appears that oscillation can be promoted and HF garbage sent to the output if bypassing is done incorrectly, or not at all if Eva is to be believed.
Go find his/her posts and extensive test results/graphs/diagrams.
Then make up your own mind.
O.K., Thanks. Kinda what I figured...
"" ...............................................................................................The concept of 'self-resonance' in this context is flawed thinking, and leads to silly designs (100nF caps in parallel with 10,000uF electros for example) that generally achieve nothing useful, other than using more components.."
I guess I will leave it the way adcom designed it, and file this away with the notion of expensive speaker cables.
"" ...............................................................................................The concept of 'self-resonance' in this context is flawed thinking, and leads to silly designs (100nF caps in parallel with 10,000uF electros for example) that generally achieve nothing useful, other than using more components.."
I guess I will leave it the way adcom designed it, and file this away with the notion of expensive speaker cables.
digital desire said:
These small capacitors are pretty cosmetics and essentially useless in normal working condition. In abnormal situation they can also led the output stages to plain faults due to transient oscillation on power supplies which may trigger SOAR and overvoltage problem in bipolar output stages and plain destructive self oscillation in MOSFET output stages...
At my knowledge, the only real improvement on power supply capacitors which may improve sound quality is reducing ESR and ESL of capacitors itself. This may be reached with proper selected components (using the same capacitors employed in switching PSU) or splitting the existing required capacitors in more components with lower nominal values but with roughly the same phisical size).
By the way: the main improvement obtained with "increasing" capacitors values (provided that they don't kill the rectifier...) is just a (relatively) reduced ESR and, in my opinion, is this reduction that "sound better" because assure, also at high frequencies (above a couple of kHz), a better "closure" in AC of the rails to the ground, so diminishing the risk of intermodulation at high audio frequencies, just where the efficiency of NFB (and of the PSRR) start to fail.
Other improvements in low frequency audio, especially when original capacitors were of adequate value (roughly a 1 or 2 mF for each ampere of output capability of *single* channel), and - not to be forget - ground connections are properly designed, are, in my opinion, somewhat questionables.
Hi
Piercarlo
"Can someone explain how this does anything?"
Since none of the big brains can think of a reason why adding small caps in the range of 10µF~100µF to the top of the 15,000µF caps in the Adcom GFA555 makes a hugh difference in how the bass sounds, I guess I would have to say it's 'magic'.
"How does this improve bass as was purported?"
As above, I guess I don't really know. Big computer grade caps just sound 'slow'. Paralleling them with small values (as above) makes them sound 'fast'.
Here is a clue though:
I used to service a brand of amplifier that was regarded as having a lot of 'punch'. When the owners would bring them in claiming they had lost their 'punch', I would find a bunch of 10µF power supply bypass caps (a pair per channel) rolling around inside the chassis. The cans had blown themselfs clean off the board. This would imply an enormous amount of current going through them, and not through the much large main filter caps.
The same thing happens with my Leach. The cans don't blow off, but the two pair of 22µF go open on a regular bassis (every few years).
This behavior suggests that on transients that the small caps must be providing large currents, despite what the EEs proclaim.
The change in sound is so dramatic that some will not like it. I had someone that owned a Bryston 3B that was in this camp, he liked the stock, muddy bass without the bypass caps. He listed mainly to RAP music though.
The small film caps on top of the big cans in the Adcom are cosmetic, they do nothing. Any effect they have is negated by the inductance in the wiring to the output stages. It is very dangerous to try and add small values of film caps (below 1µF in this context) as close as possible to the outputs. Oscillation is a very real possibility.
One last story to relate:
I modified a McIntosh for a recording studio at Cornell. The head of the music department and also the head of the studio came and talked to me after a few days of listening to the 'new' amp. "You added too much bass and treble boost" was their response. I showed them that it was ruler flat from 20hz~20Khz. I played them some real good recordings from Shefield Lab, and they compared the sound vs five other amps reputed to be good for studio use.
At that point they asked: "Why does so much of recorded music sound so bad?". My answer was: "Because their monitors and speakers can't tell them what is being recorded, and how bad their entire mic>mixer>recorder chain really is.
The quality of their studio recordings continued to improve as the speaker and mixer were worked on.
I suggest you spend $5~$10 on some caps and find out for yourself.
Since none of the big brains can think of a reason why adding small caps in the range of 10µF~100µF to the top of the 15,000µF caps in the Adcom GFA555 makes a hugh difference in how the bass sounds, I guess I would have to say it's 'magic'.
"How does this improve bass as was purported?"
As above, I guess I don't really know. Big computer grade caps just sound 'slow'. Paralleling them with small values (as above) makes them sound 'fast'.
Here is a clue though:
I used to service a brand of amplifier that was regarded as having a lot of 'punch'. When the owners would bring them in claiming they had lost their 'punch', I would find a bunch of 10µF power supply bypass caps (a pair per channel) rolling around inside the chassis. The cans had blown themselfs clean off the board. This would imply an enormous amount of current going through them, and not through the much large main filter caps.
The same thing happens with my Leach. The cans don't blow off, but the two pair of 22µF go open on a regular bassis (every few years).
This behavior suggests that on transients that the small caps must be providing large currents, despite what the EEs proclaim.
The change in sound is so dramatic that some will not like it. I had someone that owned a Bryston 3B that was in this camp, he liked the stock, muddy bass without the bypass caps. He listed mainly to RAP music though.
The small film caps on top of the big cans in the Adcom are cosmetic, they do nothing. Any effect they have is negated by the inductance in the wiring to the output stages. It is very dangerous to try and add small values of film caps (below 1µF in this context) as close as possible to the outputs. Oscillation is a very real possibility.
One last story to relate:
I modified a McIntosh for a recording studio at Cornell. The head of the music department and also the head of the studio came and talked to me after a few days of listening to the 'new' amp. "You added too much bass and treble boost" was their response. I showed them that it was ruler flat from 20hz~20Khz. I played them some real good recordings from Shefield Lab, and they compared the sound vs five other amps reputed to be good for studio use.
At that point they asked: "Why does so much of recorded music sound so bad?". My answer was: "Because their monitors and speakers can't tell them what is being recorded, and how bad their entire mic>mixer>recorder chain really is.
The quality of their studio recordings continued to improve as the speaker and mixer were worked on.
I suggest you spend $5~$10 on some caps and find out for yourself.
djk said:
This is very good advice that we should all follow.
And come back here with some answers, if possible.
The figure you mentioned is also very logic, so we can still stay within good parts and not go crazy on stupid prices.
Just a warning: if the amp is well designed we should hear very little differences.
Re: Re: Bypass caps on power supply
Ok you are using 1m or 2m as is 10,000uf or 20,000uf? is that correct?
So a 200 watt per channel @ 8 ohm amp = 5 amps RMS, then By your statement above, you are saying that at least 50,000 to 100,000uf should be used per channel? and double that for 4 ohm ratings?
I dont know any MFG's that follow that.
Zc
Piercarlo said:
Ok you are using 1m or 2m as is 10,000uf or 20,000uf? is that correct?
So a 200 watt per channel @ 8 ohm amp = 5 amps RMS, then By your statement above, you are saying that at least 50,000 to 100,000uf should be used per channel? and double that for 4 ohm ratings?
I dont know any MFG's that follow that.
Zc
For better or worse, here in the US, both uF and mfd are abbreviations for microfarad. Although technically correct, I've never seen anyone here use milliFarad, only Farad with prefixes micro, nano (fairly rare), and pico.
A few comments on the bypassing thing:
--I've been bypassing power supply caps for thirty years on scads of different amp designs. Not once, ever, have I had an amp break into oscillation. Nor has anyone that I know, nor even a friend of a friend of a friend. Nor any of the commercial designs, whether solid state or tube. Is it theoretically possible? I suppose so. It is a realistic fear? My response to that would be rendered as a string of "*" by the weenies in charge, so I won't bother wasting the keystrokes.
--It's inevitable. The if-you-can't-measure-it-it don't-exist folks get the price issue confused with the "does it make a difference?" issue. The more it costs, the more they dig in their heels and insist that it can't make a difference. Treat the two as separate questions. First, does it make a difference? Second, if it does, do you feel it's worth the asking price? Sometimes I hear things, but feel that the price is too high. So what? Does that mean I erase the memory of hearing something? No. I just didn't feel that it was worth what they were asking. Simple.
--It's not necessarily that expensive to try film cap bypasses. Panasonic has some 10uF 100V metallized polyester caps. Digikey carries them. They're not really very expensive. As film caps go, they're not all that great; decent bass, but the sound deteriorates as you go up the spectrum. Still, they're noticeably better than electrolytics. Try them in the power supply or in the circuit itself. Can you hear a difference? Yes or no. If yes, then experiment with other caps if the mood takes you. (Hint: WIMAs are better and not too terribly expensive if you can find the value you want.) Which leads me to...
--The notion that the better the amp is the less you'll hear is incorrect. I wouldn't expect to hear much difference on a table radio. As the equipment gets better, the more obvious the differences become. There's an implied insult here--or at least the possibility of one--in that no one wants to admit that his system is junk. So rather than put it to the test, some folks will simply decide that there's nothing to be heard, hence sidestepping the trouble of getting out the soldering iron, listening, and perhaps...horrors! hearing nothing (or worse...hearing something good). Omigod! Rather than admit that their DIY speakers sound like elephant flatulence, they duck the matter entirely.
--Much is made is people who "expect" to hear a difference...and so they do. Always with the implication that they are weak-minded fools who are easily lead. Never--perish the thought!--is it admitted that those who do not wish to hear a difference will not hear (or admit to hearing--after all, who wants to admit to being "weak-minded") a difference.
Nay-sayers will always find a way to stop people from trying things...always "for their own good," of course. Thumb your nose at them. Try a few caps. See what difference they make, if any. If you like what you hear, then march onwards. Otherwise, either improve your system or say in a straight-forward manner that under thus-and-such conditions that you didn't hear anything. It's much more intellectually honest than assuming that they can't make a difference.
Grey
A few comments on the bypassing thing:
--I've been bypassing power supply caps for thirty years on scads of different amp designs. Not once, ever, have I had an amp break into oscillation. Nor has anyone that I know, nor even a friend of a friend of a friend. Nor any of the commercial designs, whether solid state or tube. Is it theoretically possible? I suppose so. It is a realistic fear? My response to that would be rendered as a string of "*" by the weenies in charge, so I won't bother wasting the keystrokes.
--It's inevitable. The if-you-can't-measure-it-it don't-exist folks get the price issue confused with the "does it make a difference?" issue. The more it costs, the more they dig in their heels and insist that it can't make a difference. Treat the two as separate questions. First, does it make a difference? Second, if it does, do you feel it's worth the asking price? Sometimes I hear things, but feel that the price is too high. So what? Does that mean I erase the memory of hearing something? No. I just didn't feel that it was worth what they were asking. Simple.
--It's not necessarily that expensive to try film cap bypasses. Panasonic has some 10uF 100V metallized polyester caps. Digikey carries them. They're not really very expensive. As film caps go, they're not all that great; decent bass, but the sound deteriorates as you go up the spectrum. Still, they're noticeably better than electrolytics. Try them in the power supply or in the circuit itself. Can you hear a difference? Yes or no. If yes, then experiment with other caps if the mood takes you. (Hint: WIMAs are better and not too terribly expensive if you can find the value you want.) Which leads me to...
--The notion that the better the amp is the less you'll hear is incorrect. I wouldn't expect to hear much difference on a table radio. As the equipment gets better, the more obvious the differences become. There's an implied insult here--or at least the possibility of one--in that no one wants to admit that his system is junk. So rather than put it to the test, some folks will simply decide that there's nothing to be heard, hence sidestepping the trouble of getting out the soldering iron, listening, and perhaps...horrors! hearing nothing (or worse...hearing something good). Omigod! Rather than admit that their DIY speakers sound like elephant flatulence, they duck the matter entirely.
--Much is made is people who "expect" to hear a difference...and so they do. Always with the implication that they are weak-minded fools who are easily lead. Never--perish the thought!--is it admitted that those who do not wish to hear a difference will not hear (or admit to hearing--after all, who wants to admit to being "weak-minded") a difference.
Nay-sayers will always find a way to stop people from trying things...always "for their own good," of course. Thumb your nose at them. Try a few caps. See what difference they make, if any. If you like what you hear, then march onwards. Otherwise, either improve your system or say in a straight-forward manner that under thus-and-such conditions that you didn't hear anything. It's much more intellectually honest than assuming that they can't make a difference.
Grey
one of the things that has always frutrated me about naysayers is that most of the time, the people who claim that cant hear any difference in whatever device you are speaking about, is the people who you later find out are making these statements based on there 20 year old radio shack speakers and a mid 80's receiver!
How can you make any sort of comparisons when your system does not have enough "Resolution" to allow you to hear any sort of difference at all.
That is like trying to judge one TV to another when the glasses you are wearing are so scratched, fogged and yellowed, you can barley see through them!
Zc
How can you make any sort of comparisons when your system does not have enough "Resolution" to allow you to hear any sort of difference at all.
That is like trying to judge one TV to another when the glasses you are wearing are so scratched, fogged and yellowed, you can barley see through them!
Zc
This just in, I just now bypassed the main power supply caps on my Elgar line conditioner, in two places for *double* all that so-called trouble, and bypassed the 15V opamp supplies, and bypassed the output AC capacitors. Bypass caps are expensive sytrene and teflons, no flim flam on this particular job. Holy cow does this baby sing! And I can't measure any difference, before to after -->whatsoever<-- down to 20mV/500ps on my scope.
Hey, and to get the naysayers further clucking, this Elgar is itself fed by another Elgar and it's powering ... wait for it ... an Apple Airport Express which is serving duty as my digital source (for the moment). I can hear all that difference I'm claiming in *this* setup.
Hey, and to get the naysayers further clucking, this Elgar is itself fed by another Elgar and it's powering ... wait for it ... an Apple Airport Express which is serving duty as my digital source (for the moment). I can hear all that difference I'm claiming in *this* setup.
Well, I'm glad there are people here that really listen.
The transparency thing is something I discussed many times before, but it ended sounding as if I was mentioning something theoretical.
Common people can hear more than they imagine: the problem is to show them what they should hear, so they will miss it the next time. It's not imagination, it's more like musical notes that should be revealed.
When veils start falling we all end up winning.
The transparency thing is something I discussed many times before, but it ended sounding as if I was mentioning something theoretical.
Common people can hear more than they imagine: the problem is to show them what they should hear, so they will miss it the next time. It's not imagination, it's more like musical notes that should be revealed.
When veils start falling we all end up winning.
Carlmart, I can't say I've heard a non-bypassed amp I've enjoyed for the long term. I've owned some decent designs including, for example, a Pass Aleph. I never got around to fiddling with that amp, but suffice it to say that, despite its interesting SE/class A character, I found it dull and hard sounding, like every other unbypassed amp I owned, which is the characteristic sound of electrolytic capacitors, or to be more precise, the sound of dielectric absorption in those godawful components.
serengetiplains said:
That was probably a characteristic of the Aleph, which not many people seem to like. SE designs also seem to be more sensitive to the quality of the parts they use, but I am not sure it was only because they were not bypassed.
If that had been the case there would have been people reporting such improvements, because many would be tempted to try them.
What I think is that bypassing may improve an already good equipment, revealing more things. I don't think they can ruin a good design.
Bypassing is perhaps something that should be tuned to each specific equipment, kind of try one part and listen, then take it out. And so on.
Carlos
Re: Re: Re: Bypass caps on power supply
Wrong: 1m = 1000 uF. With 50 Hz main (as in Europe) 1mF with usual PSU is equivalent to a series resistance of 1 Ohm (not a true resistance obviously because it don't dissipate any power; just "limit" the mean voltage output of PSU). Or, as more usual, 1mF = 1 Volt peak-to-peak ripple with 1 Ampere supplied on rails.
Hi
Piercarlo
PS - With 60 Hz as used in USA ripple and "resistance" must be scaled to a 50/60; about 0.83 the result obtained in Europe with same currents and capacitors.
Zero Cool said:
Wrong: 1m = 1000 uF. With 50 Hz main (as in Europe) 1mF with usual PSU is equivalent to a series resistance of 1 Ohm (not a true resistance obviously because it don't dissipate any power; just "limit" the mean voltage output of PSU). Or, as more usual, 1mF = 1 Volt peak-to-peak ripple with 1 Ampere supplied on rails.
Hi
Piercarlo
PS - With 60 Hz as used in USA ripple and "resistance" must be scaled to a 50/60; about 0.83 the result obtained in Europe with same currents and capacitors.
It was the electrolytics, Carlmart. I had the same experience with every amp I've owned, class D, push-pull tube, SET, SS, Emm Labs preamp, including, what I'm listening to now, an Edge NL-10, the supply in which I gutted and replaced with a (regulated) Vicor switcher feeding 10uF of teflon capacitors (no electrolytics). The result? Not just slightly better sound, way better sound---in fact, probably the best I've heard. Tells me, at least, what limitations capacitors bring a piece of audio gear.
"I am not sure i understand your statment the way it is worded. are you saying it is a bad idea to put bypass caps at the main filter caps, and that they would be better located as close to the output devices as possible???"
No, I put the 10µF~100µF bypass caps right on top of the main 15,000µF filter caps. Im saying it seems to be a waste to put smaller (sub 1µF film types) here, as Adcom does. The inductance of the long wires to the output stages negate the sub 1µF caps.
Putting them where they can do the most good is not always possible (I'll try and explain this later), in that case I just leave the sub 1µF caps out. If you must try this, try a small resistor in series with the cap. The small caps can resonate with the inductance of long wires. Putting a resistor in series with them reduces the Q of the resonance. Without test equipment you really can't tell what is going on, and a sine wave driven into a resistor may not reveal the problem. Many times you can use an AM radio tuned to a blank spot on the dial instead of a 'scope. If you hear buzzes in time with the music, the amp is bursting into oscillation on musical peaks.
(later) Technically speaking, these sub 1µF caps need to go as close as possible to where the power is consumed. The big problem is the speaker ground may not be close enough to the B+ on the output devices to make the caps fit. Any long leads or wires constitute a problem here because of their inductance. IIRC, Electrocompaniet used 0.68µF with a 1 ohm resistor in series with it on their early Ampliwire units.
No, I put the 10µF~100µF bypass caps right on top of the main 15,000µF filter caps. Im saying it seems to be a waste to put smaller (sub 1µF film types) here, as Adcom does. The inductance of the long wires to the output stages negate the sub 1µF caps.
Putting them where they can do the most good is not always possible (I'll try and explain this later), in that case I just leave the sub 1µF caps out. If you must try this, try a small resistor in series with the cap. The small caps can resonate with the inductance of long wires. Putting a resistor in series with them reduces the Q of the resonance. Without test equipment you really can't tell what is going on, and a sine wave driven into a resistor may not reveal the problem. Many times you can use an AM radio tuned to a blank spot on the dial instead of a 'scope. If you hear buzzes in time with the music, the amp is bursting into oscillation on musical peaks.
(later) Technically speaking, these sub 1µF caps need to go as close as possible to where the power is consumed. The big problem is the speaker ground may not be close enough to the B+ on the output devices to make the caps fit. Any long leads or wires constitute a problem here because of their inductance. IIRC, Electrocompaniet used 0.68µF with a 1 ohm resistor in series with it on their early Ampliwire units.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.